on 08-22-2012 3:39 PM
hi
I wonder how can I manage cluster resources from luci interface such as sap database and sap instance ?
so ı need know-how to handle failover mechanism
regards
Hi,
your question is quite unspecific, but this document may help you:
Regards,
Peter
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
I suppose you mean Enqueue Replication Service?
Yes, the purpose is to keep lock consistency in case of enqueue server failure, therefore it has to be installed on a separate host.
For my understanding you'll have one SCS instance (including enqueue server) on host a, and a ERS installation on host B. ERS will then be started separatedly (not sure about that), and will start another enqueue server if the central instance enqueue fails.
Maybe these links will help you to understand the context:
http://help.sap.com/saphelp_nw04s/helpdata/en/6d/e3a03cc24efd1de10000000a114084/frameset.htm
http://help.sap.com/saphelp_nw04s/helpdata/en/6d/6a853f11ed0617e10000000a114084/frameset.htm
http://help.sap.com/saphelp_nw04s/helpdata/en/de/cf853f11ed0617e10000000a114084/frameset.htm
Regards,
Peter
Hi
config like that :
1-ASCS instance is as resource in RHEL cluster (installed using virtual name)
2-ERS also resource in RHEL cluster (also has virtual name same as ASCS instance)
3-Db instance is also resource in RHEL cluster (installed using virtual name)
3- central instance and additional instance running now with manually start (startsap R3.....)
4- when I fail node1 all resources passing the other node in cluster.ASCS instance can follow the ERS instance.also replication is ok.I check it with ensmon.
but I came across a problem while failed node is restart.
I saw this process :/usr/sap/.........../exe/sapcontrol -nr 20 -function WaitforStopped 3600 1
when I kill this process , node is running properly
yu have any idea?
2-ERS also resource in RHEL cluster (also has virtual name same as ASCS instance) --> shouldn't ERS using different virtual name?
Server host name X(x.x.x.1):
ASCS instance -> virtual ip(x.x.x.101)-> virtual host name A
assigned as a resource with virtual host name A
Server host name Y(x.x.x.2):
ERS Instance-> virtual ip (x.x.x.102) -> virtual host name B
assigned as a resource with virtual host name B
scenario:
ASCS Service started in server X with virtual name A
ERS Service started in server Y with virtual name B (replicated shm from server X for ASCS instance).
in event of failure ASCS service is relocated along with its IP resource to server Y so it can start with the shared memory(lock table) replicated by ERS.
for more complete view you can peruse following document ( there is a sample config in its Appendix)
hi Henry
really I looked that document from rhel site before.
ı read appendix A expecially SApinsatance ERS section.
there is a ERS instance as a resource in cluster.
this instance created how? it uses a virtual name like ASCS instance and DB instance.
really ı wonder exactly ERS instance how should be installed on 2 nodes for being resource in cluster?
According to documentation:
"
follow_service(svc1, svc2, master)
Function for a construct when service svc1 and service svc2 should not be running on the same node
even after failover. And if svc1 fails it should follow svc2. Then svc2 is relocated on any other node in the
cluster.
"
I.e. ERS service should be relocated also.
Regards
Roman
If you don't like this configuration try to use SAP polling concept:
"Some HA software solutions do not offer the concept that a software package obeys another package on this package’s host. For this reason SAP has introduced the polling concept: Here, a replication server runs on every possible failover host, and does not have to be controlled by the HA software. Whether you need this or not depends on your HA software. For more information, contact your HA software partner."
Regards
Roman
You install ERS service on one node only with it name. After that it is cluster services responsibility to relocate ERS to another node in the cluster with it name (and ip address, and LVM, and fs etc.) according to your cluster configuration. In two-node cluster configuration you install ASCS instance on one node and ERS instance on another and configure follow service dependency between them
Regards
Roman
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.